Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

Michael clutched the crumpled newspaper, his knuckles white. The headline screamed across the front page: “Literary Giant Accused of Plagiarism.” Underneath, a photo of Arthur Finch. The author. His author. Michael, a simple bookstore clerk, had devoured Finch's work for years, finding solace and inspiration in the words. Now, it was alleged Finch had lifted entire passages from a struggling, unknown writer - Michael.

He slowly lowered himself into a chair, the room tilting slightly. The familiar scent of old paper and ink usually comforted him, but now it felt suffocating. He reread the accusations, each word a cold, hard slap. He tried to think, but his thoughts scattered like startled birds. He felt a deep, hollow ache spread through him. It was a physical sensation, like a vacuum had been created inside his chest.

“No,” he finally breathed, the word a small, lost sound.

Emotion: dumbstruck

Cluster: Surprise / Confusion
PC1 (Valence): -1.99 Negative
PC2 (Disposition): -0.71

Role in Research

This story is one of 1,000 stories generated for the emotion dumbstruck. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 dumbstruck stories, after denoising with neutral dialogue baselines, produces the dumbstruck emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the dumbstruck vector is projected through the unembedding matrix.

Promoted:
будто0.326
не0.285
😶0.275
仿佛0.267
依旧0.258
Suppressed:
enthusi-0.278
la-0.264
K-0.260
wholeheartedly-0.255
쾌-0.254